Research On Low-overhead Integrity Verification Protocol For Dynamic Cloud Storage

With the development of cloud computing technology,cloud storage is increasingly favored by more and more individual or enterprise users.Although cloud storage has advantages such as reliability,economy,and scalability,the integrity of outsourced data in CSPs still faces multiple security threats.In order to enable users to obtain the status of their outsourced data in a timely manner,various cloud storage outsourcing data integrity verification protocols have been proposed.How to design a more secure and computationally efficient cloud storage auditing protocol has also become a research hotspot in academia.According to the current problems of storage overhead,verification efficiency,and security of cloud storage integrity verification protocols under different scenarios,this paper studies the cloud storage auditing protocols under the scenarios of single-replica cloud storage and multi-replica cloud storage.The main contributions are as follows:(1)For the current problem that the storage overhead of current cloud storage auditing protocols is large,the computational complexity of cryptographic algorithms is high,and the efficiency of data structures that support data dynamic operations is low in single-replica cloud storage scenarios,a dynamic cloud storage integrity verification protocol based on algebraic coding is proposed.The protocol includes two architectures: private verification and public verification.By using algebraic coding instead of cryptographic signature algorithms to calculate data tags to reduce computational overhead,only encoded data is stored in CSPs to reduce storage overhead and effectively protect user privacy data.A dynamic index skip list data structure is proposed to achieve efficient data dynamic operations.In addition,according to the security definition of cloud storage auditing protocols,the correctness,security and privacy protection of the proposed protocol are proved.Finally,through theoretical analysis and experimental evaluation,the advantages of this scheme in privacy protection,storage overhead,communication overhead,computational overhead and data dynamic efficiency are verified.(2)For the problem that it is difficult to recover data after data loss in single-replica cloud storage scenarios and CSPs face risks such as compressed storage and service interruption in multi-CSP multi-replica cloud storage scenarios,a dynamic auditing protocol for multi-CSP multi-replica cloud storage based on CSP reputation scores is proposed.The protocol distinguishes different replicas of data using random masking technology according to CSP identity and stores different replicas of data in different CSPs.Even if a CSP stops service,users can still download data replicas from servers of other CSPs to restore original data.Homomorphic linear authenticator technology is used to support independent verification and aggregation verification of integrity evidence.According to CSP reputation scores,challenges are focused on CSPs with low reputation scores to improve damaged data detection rate.A high-efficiency error location and data recovery strategy based on CSP reputation scores and independent/aggregation verification technology is proposed.In addition,in order to support data dynamics,MHT is improved into multi-replica MHT data structure that supports multi-replica data storage.The hash values of all replicas of the same data block are used as sibling nodes to reduce the height of ordinary multi-replica MHT.Based on the security model designed for multi-replica cloud storage protocols,the correctness and security of the protocol are proved.Finally,through theoretical analysis and experiments,the advantages of this scheme in integrity verification and error location and data recovery are verified.